Open latenitefilms opened 1 year ago
I upgraded. FCP 10.6.5 wouldnāt complete the upgrade of the library and wouldn't open after several attempts and force quits of the app, then the computer. Eventually the library opened and started the update process, but would repeatedly hang. Why, I donāt know. I would just watch as it has to update 112 items in the library and I donāt have 112 projects btw. Fortunately I backed up 10.6.5, that library, and the cache folder. 10.6.6 seems to have had many problems to many problems so I didnāt update and decided to wait for 10.6.7. This same library is finally working (as in opening) under 10.6.7, but it was a rough start.
Computer setup: Iām on an M2 Max MacBook Pro with 64GB of memory, clamshell mode with two monitors, but only my 34" LG Ultrwaide was plugged in at the time along with my CalDigit TS4 Thunderbolt Dock. I am using macOS Ventura 13.4.1 (c) (22F770820d)
I may go back to 10.6.5 since I have everything backed up. We'll see how things go.
Chris you're already very familiar with my library. I haven't upgraded to BRAW Toolbox 1.3.0 yet. We're at locked picture and I have everything backed up for going back to 10.6.5, so I thought 10.6.7 would be alright.
crash experienced while using the Text Inspector on a Title Template. the timeline is only about 30s and contains audio & a few title clips; no video. [M1 Max, 64GB] [macOS 13.5] [FCP 10.6.7]
Any chance you could share the full crash log? You can find it in:
~/Library/Logs/DiagnosticReports/
You can then share the full contents as a GitHub Gist.
Any chance you could share the full crash log? You can find it in:
~/Library/Logs/DiagnosticReports/
You can then share the full contents as a GitHub Gist.
https://gist.github.com/sharedscene/85c21d0b441356402c9f32ca163f607a
Legend, thanks!
Here's the crashed thread:
Thread 45 Crashed:: Dispatch queue: com.apple.flexo.playertokq.sched
0 ProGraphics 0x10f23ec88 PGPerThreadSetCurrentContextSentry::PGPerThreadSetCurrentContextSentry() + 24
1 Ozone 0x113c653ec OZFxPlugSharedBase::scheduleTokens(OZRenderParams const&, std::__1::vector<std::__1::pair<CMTime, std::__1::vector<CMTime, std::__1::allocator<CMTime>>>, std::__1::allocator<std::__1::pair<CMTime, std::__1::vector<CMTime, std::__1::allocator<CMTime>>>>> const&, PMFrameRequest&, std::__1::vector<FFScheduleToken*, std::__1::allocator<FFScheduleToken*>>*) + 104
2 Ozone 0x113875538 OZScene::scheduleTokens(OZRenderParams const&, std::__1::vector<std::__1::pair<CMTime, std::__1::vector<CMTime, std::__1::allocator<CMTime>>>, std::__1::allocator<std::__1::pair<CMTime, std::__1::vector<CMTime, std::__1::allocator<CMTime>>>>> const&, std::__1::vector<FFScheduleToken*, std::__1::allocator<FFScheduleToken*>>*) + 528
3 MotionEffect 0x10f35bb2c -[FFMotionEffect newEffectSpecificTokensAtTime:duration:withInputStream:context:schedInfo:downstreamPT:] + 1048
4 Flexo 0x107c0bf50 -[FFStreamVideoEffect newScheduleTokenAtTimeIgnoringCache:duration:context:schedInfo:downstreamPT:] + 1584
5 Flexo 0x107c34488 -[FFStreamVideo newScheduleTokenAtTime:duration:context:schedInfo:downstreamPT:] + 564
6 Flexo 0x107c0ba54 -[FFStreamVideoEffect newScheduleTokenAtTimeIgnoringCache:duration:context:schedInfo:downstreamPT:] + 308
7 Flexo 0x107c34488 -[FFStreamVideo newScheduleTokenAtTime:duration:context:schedInfo:downstreamPT:] + 564
8 Flexo 0x107c0bcbc -[FFStreamVideoEffect newScheduleTokenAtTimeIgnoringCache:duration:context:schedInfo:downstreamPT:] + 924
9 Flexo 0x107c34488 -[FFStreamVideo newScheduleTokenAtTime:duration:context:schedInfo:downstreamPT:] + 564
10 Flexo 0x108252200 -[FFAnchoredObjectSegmentedVideoStream newScheduleTokenAtTimeIgnoringCache:duration:context:schedInfo:downstreamPT:] + 256
11 Flexo 0x107c34488 -[FFStreamVideo newScheduleTokenAtTime:duration:context:schedInfo:downstreamPT:] + 564
12 Flexo 0x108484c20 -[FFMCSwitcherVideoStream newScheduleTokenAtTimeIgnoringCache:duration:context:schedInfo:downstreamPT:] + 480
13 Flexo 0x107c34488 -[FFStreamVideo newScheduleTokenAtTime:duration:context:schedInfo:downstreamPT:] + 564
14 Flexo 0x108014e5c -[FFPlayerScheduledData _scheduleInternal] + 1348
15 Flexo 0x1084e6f9c FFApplicationHandleExceptionThatShouldNotBeIgnoredInBlock + 20
16 Flexo 0x108761158 -[FFPlayerSchedTokenQueue _performScheduling] + 268
17 libdispatch.dylib 0x1a7e2e874 _dispatch_call_block_and_release + 32
18 libdispatch.dylib 0x1a7e30400 _dispatch_client_callout + 20
19 libdispatch.dylib 0x1a7e37a88 _dispatch_lane_serial_drain + 668
20 libdispatch.dylib 0x1a7e3862c _dispatch_lane_invoke + 436
21 libdispatch.dylib 0x1a7e43244 _dispatch_workloop_worker_thread + 648
22 libsystem_pthread.dylib 0x1a7fdc074 _pthread_wqthread + 288
23 libsystem_pthread.dylib 0x1a7fdad94 start_wqthread + 8
Looks like there's an issue in the ProGraphics
framework related to the number of threads.
another one so far today with the same heading detail for the thread:
Thread 72 Crashed:: Dispatch queue: com.apple.flexo.playertokq.sched
listing ProGraphics
framework at the top
https://gist.github.com/sharedscene/dc441c80643bdac8a8d4880263d7e715
Hello fellow FCPers! Just wanted to share that I too have been having problems with crashing. It's usually just a minor annoyance if it crashes because everything's been saved. I hope more and more people complain to Apple about this so that it gets fixed.
The media files are stored on Samba shares on an Ubuntu NAS using the latest MacOS-specific extensions in my smb.conf.
Mac Studio M1 Max 64gb Ventura 13.5 Latest Pro Video codecs installed No plugins whatsoever
One crash log found here (some apparently weren't saved?): https://gist.github.com/jkominkiewicz/b84faa82319efe96ac82da52886bf56b
Thread 46 Crashed:: Dispatch queue: com.apple.helium.rq.gpu-ru1.vs0.gcd
0 ProGraphics 0x110c5ec88 PGPerThreadSetCurrentContextSentry::PGPerThreadSetCurrentContextSentry() + 24
1 TextFramework 0x112dc0ebc TXHeliumNode::preRenderGlyphsMetal(HGPage*) + 108
2 TextFramework 0x112dc3614 TXHeliumNode::RenderPageMetal(HGPage*) + 76
3 Helium 0x104c96ba4 HGGPURenderer::RenderPage(HGNode*, HGPage*) + 832
4 Helium 0x104c96650 HGGPURenderer::ProcessOnePage(HGNode*, HGRect, HGROIsInfo&, HGFormat, HGBuffer*, bool, bool) + 720
5 Helium 0x104c95784 HGGPURenderer::ProcessNodeInternal(HGNode*, HGRect, HGFormat, HGBuffer*, bool, bool, bool) + 460
6 Helium 0x104c92f28 HGGPURenderer::GetNodeBitmap(HGNode*, HGRect, HGBuffer*, bool) + 436
7 Helium 0x104da76f0 HGNode::PullMetalTextures(HGPage*, int) + 400
8 Helium 0x104da76a4 HGNode::PullMetalTextures(HGPage*, int) + 324
9 Helium 0x104da76a4 HGNode::PullMetalTextures(HGPage*, int) + 324
10 Helium 0x104da76a4 HGNode::PullMetalTextures(HGPage*, int) + 324
11 Helium 0x104da76a4 HGNode::PullMetalTextures(HGPage*, int) + 324
12 Helium 0x104da76a4 HGNode::PullMetalTextures(HGPage*, int) + 324
13 Helium 0x104da76a4 HGNode::PullMetalTextures(HGPage*, int) + 324
14 Helium 0x104da76a4 HGNode::PullMetalTextures(HGPage*, int) + 324
15 Helium 0x104da76a4 HGNode::PullMetalTextures(HGPage*, int) + 324
16 Helium 0x104da7f08 HGNode::RenderPageMetal(HGPage*) + 76
17 Helium 0x104c96ba4 HGGPURenderer::RenderPage(HGNode*, HGPage*) + 832
18 Helium 0x104c96650 HGGPURenderer::ProcessOnePage(HGNode*, HGRect, HGROIsInfo&, HGFormat, HGBuffer*, bool, bool) + 720
19 Helium 0x104c95784 HGGPURenderer::ProcessNodeInternal(HGNode*, HGRect, HGFormat, HGBuffer*, bool, bool, bool) + 460
20 Helium 0x104c92648 HGGPURenderer::RenderTiles(HGNode*, HGRect, HGFormat, HGBitmap*, bool, bool) + 268
21 Helium 0x104d658f4 HGRenderer::RenderNode(HGNode*, HGRect, HGFormat, HGBitmap*) + 144
22 Helium 0x104d69298 HGRenderer::RenderNodes(std::__1::vector<HGRendererOutput, std::__1::allocator<HGRendererOutput>> const&, bool) + 924
23 Helium 0x104d68ac8 HGRenderer::RenderNode(HGRendererOutput const&, bool) + 100
24 Helium 0x104ca69b8 HGRenderExecUnit::ExecuteRenderGPU(HGRenderJob*) + 1216
25 libdispatch.dylib 0x184907518 _dispatch_block_async_invoke2 + 148
26 libdispatch.dylib 0x1848f8400 _dispatch_client_callout + 20
27 libdispatch.dylib 0x1848ffb68 _dispatch_lane_serial_drain + 892
28 libdispatch.dylib 0x18490062c _dispatch_lane_invoke + 436
29 libdispatch.dylib 0x18490b244 _dispatch_workloop_worker_thread + 648
30 libsystem_pthread.dylib 0x184aa4074 _pthread_wqthread + 288
31 libsystem_pthread.dylib 0x184aa2d94 start_wqthread + 8
Thanks @jkominkiewicz! That's consistent with @sharedscene's crashes too. Something is broken in Motion.
So, since it's related to Motion, I looked to see whether I had updated to the latest version. It turns out I was one release behind since having updated to the newest FCP release. I've updated to ver. 5.6.5 and I haven't had a crash in the last half hour. Fingers crossed.
UPDATE: So, after having used it for around four hours with the updated Motion: Crashes have stopped, but I'm now getting the frozen viewer requiring a restart of FCP and the warning that my computer can't handle 4K footage.
Here are some comments based on study of framework differences between 10.6.6 and 10.6.7, also several 10.6.7 crash logs. I used the MacOS command-line 'nm' utility to study the framework symbol differences between 10.6.6 and 10.6.7, with emphasis on the crashing call stacks people have reported.
The crash usually happens in the ProGraphics framework, which is written in C++. While much of FCP is written in Objective-C, a non-trivial amount is written in C++. The crashing method is called a 'destructor', which cleans up resources for an object being deallocated.
Doing a 'diff' on the symbols between the two versions of ProGraphics framework, there were possibly a lot of changes. In the 10.6.6 version many methods were had '.cold.1', '.cold.2', etc. suffixes. That indicates Profile-Guided Optimization had been run, which pulls out and repacks less-executed code blocks so the remaining 'hot' code blocks obtain better hit ratios on the CPU instruction cache. The fact that many of the '.cold' blocks vanished between 10.6.6 and 10.6.7 could imply the underlying source code was modified (IMO more likely), or it might mean a different profile exercise was done.
The crashing PGPerThreadSetCurrentContextSentry is one of those which changed between 10.6.6 and 10.6.7: https://photos.smugmug.com/photos/i-3Lmxhmj/0/398ab6ff/X5/i-3Lmxhmj-X5.jpg
Also the code in the 'Ozone' framework (which contains the Motion runtime engine that implements FCP effects) had lots of apparent changes. This is just speculation, but maybe some of those source code or profiling changes were related to shoe-horning FCP into iPadOS. See similar diff comparison of Ozone symbols between 10.6.6 and 10.6.7: https://photos.smugmug.com/photos/i-hKpgbrn/0/4ce40842/X5/i-hKpgbrn-X5.jpg
Yes, I compared the ProGraphics
framework between Final Cut Pro 10.6.6 and 10.6.7 using Hopper, and there's definitely some differences in PGPerThreadSetCurrentContextSentry
. It'll be interested to see what the release notes are for the next update, but I'm assuming it'll just be "general stability improvements" or similar.
After update, Voice Isolation box not checkable. Freezing viewer another problem which can be temporarily fixed by trashing preferences but it recurs.
NOTE: If you've come here via a social media link, this thread is part of FCP Cafe's Bug Tracker.
You can learn more here: https://fcp.cafe/bugtracker/
FINAL CUT PRO 10.6.8 IS OUT NOW!
If you're currently using Final Cut Pro 10.6.6 or earlier we recommend sticking with that for the time being.
However, if you're already using Final Cut Pro 10.6.7 we recommend updating, as it resolves some issues with the Viewer going black and various crashes.
Please make sure you make a backup of your Final Cut Pro application by creating a ZIP file of it, before updating. You should also manually backup any libraries before updating.
INTRODUCTION:
There's a lot of people out there having serious problems with Final Cut Pro 10.6.7, and yet, there's others that say they haven't had any issues yet. To try and find out what's going on, the purpose of this thread is to document all the different scenarios, so we can see if there's a common cause - for example, maybe it only happens on Apple Silicon or Intel, or maybe it only happens on older versions of macOS.
Dylan Bates - The Final Cut Bro has posted a video here explaining.
DISCUSSIONS ON OTHER FORUMS/GROUPS:
TIPS TO TRY AND MAKE IT MORE STABLE:
If you're still having issues, please let us know in this thread, and include your macOS version and Mac specs.
WORKAROUNDS:
We recommend sticking with Final Cut Pro 10.6.5 or Final Cut Pro 10.6.6 for now. Whilst there are SOME users that are reporting that 10.6.7 is stable, there's not enough improvements in that version to take the risk.
You can manually copy and paste the older Final Cut Pro application from a backup or another machine to your Mac. Final Cut Pro is a universal app, so the same binary works on both Intel and Apple Silicon.
Inside the Final Cut Pro application bundle is your unique App Store receipt - therefore it's only recommended use share your Final Cut Pro application with machines that are using the same Apple ID, otherwise it might trigger Apple's copy protection processes.
If you don't have a copy of Final Cut Pro 10.6.5 handy, if you have an older machine running macOS Big Sur, you can download Final Cut Pro 10.6.5 on it. You can download a copy of macOS Big Sur here and install it on an external SSD.
If you're already updated a library to Final Cut Pro 10.6.7's format, Final Cut Pro will have automatically made a backup in your backups folder prior to the migration. Alternatively, you can also export out a FCPXML and import it into an older version, keeping in mind that FCPXML isn't lossless (i.e some Title parameters don't apply correctly).
BUG & CRASH TRACKER:
Below is an evolving table of known hardware/software configurations that have reported issues:
THIRD PARTY PLUGINS:
It's important to note that since Final Cut Pro 10.6.6 Apple has dropped FxPlug3 support.
In Final Cut Pro 10.6.5 and earlier, FxPlug3 plugins ran "inside" Final Cut Pro, and a bad plugin could crash Final Cut Pro.
However, since Final Cut Pro 10.6.6, only FxPlug4 is supported, and all FxPlug4 plugins are run "out of process" - meaning they're effectively their own application, and if they crash, Final Cut Pro doesn't. Generally speaking if a FxPlug4 plugin crashes, then Final Cut Pro will just try restart it, and if it crashes 5 times or so, then it'll display the "this plugin is not responding" message. Workflow Extensions also run out of process.
What this means is that since Final Cut Pro 10.6.6, if Final Cut Pro crashes, it's NOT a third party FxPlug4 effect or a Workflow Extension - it's something in Apple's own code, and there's nothing third party developers can do about it, so there's no point emailing FxFactory, CoreMelt, MotionVFX, etc. to complain.
This is explained in more detail on FCP Cafe here.
STABLE:
BLACK SCREEN BUG (REQUIRES MANUAL RESTART):
I'm running MacBook Pro (16-inch, 2021, Apple M1 Max, 64 GB, 8TB SSD) & macOS Monterey 12.6.6. Using a LG 5K external monitor. I've definitely had to restart FCPX quite a few times because the Viewer just goes black, and I can't bring it back to life. I had one instance where adjusting audio made the viewer strobe, also requiring a manual restart.
No crashes thus far. Trashing preferences had no impact. Annoyingly, I also lost my Destination Presets when updating to 10.6.7.Richard Coppola writes:
BUGGY:
CRASHY:
I've just had my first crash - see #225.
Nguyį» n Quį»c HoĆ ng writes:
COMPRESSOR:
Steffan Andrews writes:
MOTION:
Tycondaroga100 writes: